Continental Accretion: From Oceanic Plateaus to Allochthonous Terranes

STOR

Z. Ben-Avraham; A. Nur; D. Jones; A. Cox
Science, New Series, Vol. 213, No. 4503. (Jul. 3, 1981), pp. 47-54.
Stable URL:
http://links.jstor.org/sici?sici=0036-8075%2819810703%293%3A213%3A4503%3C47%3ACAFOPT%3E2.0.CO%3B2-B
Science is currently published by American Association for the Advancement of Science.
Your use of the JSTOR archive indicates your acceptance of JSTOR' s Terms and Conditions of Use, available at
http://www.jstor.org/aboutiterms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you
have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and
you may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www .j stor .org/journals/ aaas.html.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or
printed page of such transmission.
JSTOR is an independent not-for-profit organization dedicated to creating and preserving a digital archive of
scholarly journals. For more information regarding JSTOR, please contact support@jstor.org.
http://www .j stor.org/
Wed Mar 22 12:05: 10 2006
Continental Accretion: From Oceanic
Plateaus to Allochthonous Terranes
magnitude greater than that of layer 2c in
the basement of normal oceanic crust.
Comparison of some typical oceanic and
continental velocity structures (Fig. 3)
suggests that the Ontong Java Plateau
and the Seychelles Bank could be sub-
merged continental fragments similar to
the Lord Howe Rise, or anomalously
thickened oceanic crust.
Z. Ben-Avraham, A. Nur, D. Jones, A. Cox
Magnetic lineations. Most plateaus
exhibit weak or no magnetic lineations,
suggesting that they are not formed as
typical oceanic crust.
Abundant geological and geophysical
data obtained from work on land demon-
strate tectonic complexities that seem far
greater than those of the ocean floor,
where the features are fairly well ex-
plained by plate tectonics. The tectonic
evolution is particularly complex for the
large cordilleran chains of the world.
Although some of these mountain chains
appear to be the result of classical plate
collisions, the origin of others remains
enigmatic. Many are known to be com-
posed of numerous juxtaposed slivers
with dramatically different tectonic and
stratigraphic histories. At least some
slivers appear to have originated far
away from the stable cores of the conti-
nents. The term allochthonous terranes
describes such regions; they are tectoni-
cally and stratigraphically distinct from
adjacent regions, separated from adja-
cent terranes by bounding faults, and
came to their present resting places from
distant points of origin. For example,
large parts of the mountainous regions of
western North America are composed of
such allochthonous terranes, some of
which have migrated thousands ofkilom-
eters, to be added by accretion to the
western continental United States, Can-
ada, Alaska, Siberia, and other parts of
the Pacific rim (1).
We suggest that modern analogs of
many allochthonous terranes may be
found in the oceans, in the puzzling
topographic ridges, rises, or plateaus
present on the ocean floor. We believe
that some of the oceanic plateaus, which
comprise about 10 percent of the ocean
floor, are modern allochthonous terranes
in migration, moving with the oceanic
plates in which they are embedded and
fated eventually to be accreted to conti-
nents adjacent to the subduction zones
that ring the Pacific. The plateaus in the
oceans and the allochthonous terranes
on land may provide one of the major
missing links in geodynamics: the link
between hypotheses of plate tectonics in
oceans and accretion tectonics in the
continents.
SCIENCE, VOL. 213, 3 JULY 1981
What Are Oceanic Plateaus?
Oceanic plateaus are anomalously
high parts of the sea floor that are not at
present parts of continents, active vol-
canic arcs, or active spreading ridges.
Included are rises that have been de-
scribed as extinct arcs (2), extinct
spreading ridges, detached and sub-
merged continental fragments (3), anom-
alous volcanic piles (4), or uplifted oce-
anic crust. Figure I shows the locations
Gravity. Generally, the plateaus do
not exhibit significant isostatic anoma-
lies, implying more or less complete
compensation.
Nature of margins. Various types of
plateau margins have been identified,
such as an ancient subduction zone at
the northern margin of the Bowers Ridge
(6) and a rifted margin at the eastern edge
of the Ontong Java Plateau (7). The
nature of most plateau margins, howev-
er, is not known.
Surface geology, drilling, and dredg-
ing. Several plateaus show strong conti-
Summary. Some of the regions of the anomalously high sea-floor topography in
today's oceans may be modern allochthonous terranes moving with their oceanic
plates. Fated to collide with and be accreted to adjacent continents, they may create
complex volcanism, cut off ahd trap oceanic Crust, and cause orogenic deformation.
The accretion of plateaus during subduction of oceanic plates may be responsible for
mountain building comparable to that produced by the collision of continents.
of more than 100 present-day oceanic pla-
teaus. Although particularly abundant in
the western Pacific (5) and Indian
oceans, they are found also in the Atlan-
tic Ocean, the Caribbean Sea, and the
Mediterranean Sea. Many of the large
oceanic plateaus exhibit several common
characteristic features.
Morphology. Most plateaus rise thou-
sands of meters above the surrounding
sea floor. Some, such as the Seychelles
Bank, rise above sea level, whereas oth-
ers, such as the Ontong Java Plateau; are
1500 to 2000 meters below sea level.
Crustal structure. Most of the plateaus
for which seismic refraction and gravity
data are available have estimated crustal
thicknesses ranging from 20 to more than
40 kilometers, which are two to five
times the thickness of usual oceanic
crust (- 8 km) (Fig. 2).
Crustal velocities. Some plateaus have
an upper crust 5 to IS km thick, where
compressional wave velocities are in the
range 6.0 to 6.3 km per second. This is
typical not only of one of the layers of
oceanic crust (layer 2c), but also of gra-
nitic rocks in the continental crust. The
thickness of this layer is an order of
nental affinities. For example, Precam-
brian granitic basement is exposed in the
Seychelles Islands in the middle of the
Indian Ocean. Granitic basement was
found in the Parace! Islands in the South
China Sea (8). Dredging of the Agulhas
Plateau yielded Precambrian or Paleozo-
ic granitic rocks (9). These observations
suggest that parts of these plateaus are
submerged continental fragments.
Other plateaus are of volcanic origin.
For example, the Cocos and Carnegie
ridges appear to be the result of a contin-
uously active hot spot that extruded ba-
saltic rocks onto the overriding Cocos
plate (10).
Drilling into the Ontong Java Plateau
revealed a few meters of Early Creta-
ceous basalt beneath more than 1 km of
calcareous sediments, indicating shallow
deposition since Early Cretaceous time
(11). The nature of the rock underlying
the Ontong Java volcanics is not known.
Little is known about the composition at
depth of most other plateaus.
Z. Ben-Avraham, A. Nur, and A. Cox are with the
Department of Geophysics, Stanford University,
Stanford, California 94305. D. Jones is with the U.S.
Geological Survey, Menlo Park, California 94025.
0036-8075/8110703-0047$02.0010 Copyright 1981 AAAS 47
7<;
60
0
60
CAPE
RISE


. if
() 0'"'.-"
!U @

CROZET

SEAMOUNTS
70 r ----=", /, I '"=vi 'P ./ I u'p >-- "Sf--, c:Y = 1', ii'--=-="
200 ,_n __ n , __ n , __ n
Fig. 1. Distribution of oceanic plateaus (shaded areas) in the world's oceans,
Relative motions. Measurements of
magnetic inclination from cored sedi-
ments, from, for example, the Ontong
Java Plateau, indicate substantial migra-
tion with time (12). These data are insuf-
ficient to determine relative motion be-
tween plateaus and surrounding oceanic
crust, but seismic data suggest that little
or no slip occurs at the margins, except
perhaps during periods of collision. We
assume, therefore, that most plateaus
are moving with the oceanic plates in
which they are embedded.
Consumption of Oceanic Plateaus
A small number of plateaus are being
consumed at subduction zones, with pro-
found geological effects that include re-
duced seismicity and shifts in volcanic
activity (13). Furthermore, the distribu-
tion of oceanic plateaus (Fig. 1) suggests
that new collisions will occur; for exam-
ple, the Shatsky Rise may collide with
Japan (Fig. 4). Consequently, it is rea-
sonable to assume that the consumption
of oceanic plateaus at plate boundaries
was an important tectonic process in the
past.
Eastern Pacific. Collisions between
oceanic plateaus and subduction zones
are occurring in the southeastern Pacific,
where the Juan Fernandez, Nazca, Car-
negie, and Cocos ridges are colliding
with the western margin of South Ameri-
ca. These collisions exert a remarkable
control on the seismicity, volcanism, and
morphology on the adjacent continent
(13).
The internal structure and composi-
tion of these plateaus are not known in
detail, but they may be volcanic in ori-
gin, possibly the result of a hot spot (10,
14). Since the ridges are in isostatic
equilibrium, with deep, light roots, it is
reasonable to assume that they do not
sink with normal oceanic crust at sub-
duction plate boundaries (13, 15).
Where the Nazca Ridge, towering
more than 1500 m above the sea floor,
collides with South America, the trench
is greatly diminished in depth. For 1500
km north of this point, there is a gap in
present-day volcanism, and the dip of
the seismic plane is anomalously shallow
in this zone. A similar volcanic gap is
present farther south, just north of the
point where the Juan Fernandez Ridge
collides with the continent off the coast
of Chile. A third gap extends south of the
point at which the Cocos Ridge meets
the continent in Panama (Fig. 5). Not
only seismicity (16-19) and trench con-
figuration but also volcanic activity in
South and Central America are directly
related to the oblique collision and con-
sumption of the Juan Fernandez and
Nazca ridges carried by the Nazca plate
(13).
The oblique collision of the ridges may
be responsible for the volcanic gaps.
During subduction of normal oceanic
lithosphere, typically dipping 30 to 45,
volcanic activity is continuous. The ar-
rival of buoyant crust at the trench
causes the cessation not only 'ot subduc-
tion (20) but also of volcanism, perhaps
because of the reduction in water supply
needed for melting in the downgoing
slab. The oblique orientation of the
ridges, relative to the movement of the
Nazca plate, causes the volcanic gaps to
migrate along the plate boundary. As the
fragments of the ridges become well em-
bedded, the oceanic crust behind begins
to be subducted again, forming first a
new trench and then a new seismic slab
Fig. 2. Relief versus
crustal thickness of
several oceanic pla-
teaus. Most have
crustal thicknesses in-
termediate between
oceanic and continen-
tal values.
E
-=
~
Q;
6.0
4.0
3.0
cc 2.0
1.0
Oceanic
crust
seaward of the old one. This sequence of
events leads to an apparent flattening of
the active seismic zone (21, 22), eventu-
ally followed by renewal of volcanic ac-
tivity and normal subduction of the oce-
anic plate.
In addition to transient effects, such as
changes in volcanism during the accre-
tion of oceanic plateaus, more perma-
nent geological imprints may mark the
process. The most likely imprints are the
allochthonous terranes, many of which
are embedded in the margins of conti-
nents, particularly in those that under-
went orogenesis.
Northern and western Pacific. The
transformation of oceanic plateaus into
allochthonous terranes may take differ-
ent forms depending on, among other
things, whether tectonic stress is high or
low. Seismicity implies that stress in the
o ~ ____ J - L - ~ ____ ~ __ ~ __ ~ ______ ~ ______ -L __ -L
o 10 20 30 40 60
Depth to Moho (km)
Fig. 3. Comparison of
some typical oceanic
and continental crust-
OJ
>
al structures (36). Al- " '"
OJ
OJ
'"
~
...,
>-
though
."
~
OJ
'" morphologi- -g c g>
"
"
cally similar, the On-
OJ
OJ ()
E
E
"0
"
>-
c
a; ()
" 0
I
tong Java Plateau is
.Ii 0
'"
structurally dissimilar
0
to Iceland (69, 70) or
to typical oceanic 10
crust. It is, however,
remarkably similar to
20
a typical shield struc-
E
ture, where compres-
c
sional wave velocity ~
30
C.
in thick upper crust is
" 0
6.1 km/sec, and to 40
the Seychelles Bank,
which is known to be
continental (71). Typ-
60
ical orogenic roots
like those of the Hi- 60
malayas (72) and the
Andes (73) are even
thicker.
eastern Pacific is high; most energy re-
leased by earthquakes is measured there
(23). Because of this high-stress regime,
the disruptive effects of the consumption
of ridges, such as the Nazca and Cocos,
are only temporary, and the colliding
ridges may undergo extensive deforma-
tion during accretion. However, the con-
figuration of the plate boundary will be
changed only by a modest migration of
the trench toward the ocean basin, after
which unimpeded subduction will re-
sume.
In contrast, the plate boundary geome-
try may change significantly upon pla-
teau collision in low-stress regimes, such
as those bounding the west and north
Pacific, where the subduction zones are
adjacent to island arcs and marginal
seas, not a continent. Subduction may
reverse direction or the subduction zone
may migrate to the oceanic side of the
plateau (13, 24). The formation of several
marginal seas around the Pacific can be
explained in this way (25).
An example of a low-stress regime is
the Bering Sea, which is thought to be a
marginal sea formed by the Aleutian arc,
which trapped a portion of the Kula plate
(26). Before the formation of the Aleu-
tian arc in the late Mesozoic or early
Tertiary, subduction probably took place
along the present-day continental margin
of the Bering Sea (26). As the arc
formed, subduction shifted to the Aleu-
tian Trench some time before the change
in motion of the Pacific plate 43 million
years ago (27). This shift in subduction
can be explained by the collision of an
oceanic plateau with the Mesozoic sub-
duction zone.
At present the Bering Sea has three
135 150 165 180 165 135 150 165 180 165
Fig. 4. Sketches of possible future events in the northwest Pacific. based on present-day plate
motion parameters (74). (A) Present-day configuration of Shatsky Rise. Hess Rise, Emperor
Seamounts, and Hawaiian Ridge. (B) In 6 million years, all the plateaus will have moved to the
northwest, and the Meiji Guyot, after colliding with the subduction zone, will become part of
the Kamchatka margin. (C) In 12 million years, the Shatsky Rise will collide with north Honshu,
Hokkaido, and the Kuriles. At this stage the trench might move to the oceanic side of the
plateau and a new marginal sea, the "Shatsky Sea," could form. (D) In 18 million years, the
Shatsky Rise, Hess Rise, and Emperor Seamounts will be part of the Eurasia plate, and new
plate boundaries will form in this region.
50
large oceanic plateaus and ridges: the
Umnak Plateau, the Bowers Ridge, and
the Shirshov Ridge. Refraction data from
the Bowers Ridge (28) and the U mnak
Plateau (29) indicate that a thickened
welt of crustal material is present be-
neath both features. The Bowers Ridge,
with altered andesitic rocks, a positive
magnetic anomaly over its crest, and a
sediment wedge on its northern side, is
probably an extinct island arc. Multi-
channel seismic profiles (6) reveal that
there was a subduction zone on the
northern side of the Bowers Ridge and
that the Bering Sea margin was also a
subduction zone. In the past, the Bowers
Ridge must have moved toward the Be-
ring Sea margin.
It is not clear whether the Umnak
Plateau, now situated between the Be-
ring Sea margin and the Aleutian Ridge,
was formed in situ or not, but it is
possible that, like the Bowers Ridge, it
came from elsewhere. Thus, a possible
scenario is that before formation of the
Aleutian Ridge, the proto-Bowers Ridge
and proto-U mnak Plateau moved into
their present positions in the Bering Sea
(25). The collision of the U mnak Plateau
with the then convergent Bering Sea
margin may have caused subduction to
terminate and move southward, resulting
in the formation of the Aleutian arc.
Similarly, the Shirshov Ridge, separating
the Aleutian and Komandorsky basins,
could have been formed along a large
transform fault that was active during the
northward motion of the Kula plate (Fig.
6) or by rifting away from Kamchatka
(25). Both mechanisms can explain why
the Komandorsky Basin contains less
sediment, has higher heat flow, and thus
is probably younger than the Aleutian
Basin.
A similar process may be responsible
for the two distinct volcanic arcs around
Japan. One is the northeast Japan arc,
which includes the Kurile Islands and
the northeast Japan and Izu-Marianas
arcs; the other is the Ryukyu arc. In the
past, one continuous subduction zone
existed along the Japan arc from the
Kurile arc to the Ryukyu arc (30). It has
been suggested (31) that aseismic ridges
originally located in the south moved
north with the Kula plate and eventually
collided with the Japan arc, causing the
bend in the arc and the rotation of north-
ern Honshu. We suggest further that it
was the proto-Izu Bonin arc which came
from the south and collided with the
subduction zone. Subduction then shift-
ed to the east, and two distinct arcs were
formed, isolating parts of the Kula plate
between the Ryukyu arc and the proto-
Izu Bonin arc.
SCIENCE, VOL. 213
Allochthonous Terranes Along the
Pacific Margins
Indirect geologic evidence indicates
that plateaus similar to those that exist in
ocean basins today also existed in an-
cient ocean basins. These ancient pla-
teaus can now be recognized only by
their remanents that have been incorpo-
rated into continental masses in the form
of allochthonous terranes; their stratigra-
phy and paleomagnetism indicate distant
origins. Figure 7 shows several alloch-
thonous terranes along the northeast Pa-
cific margin that were probably oceanic
plateaus at some time.
Critical evidence for extensive migra-
tion of terranes comes from measure-
ments of magnetic inclination (Table 1),
which are used to decipher the latitudinal
component of motion. Many of the
North Pacific allochthonous terranes in
Alaska and northeast Asia show migra-
tions of several thousand kilometers
over periods of tens of millions of years,
with inferred velocities of about 5 centi-
meters per year (32). Paleomagnetic azi-
muths or declinations are commonly
anomalous, suggesting that many ter-
ranes have also undergone substantial
rotation (33). Episodes of accretion of
allochthonous terranes have been sug-
gested as an important part of the pro-
cesses of crustal growth (1), crustal
shearing (34), mountain building (35, 36),
and the creation of marginal seas (24, 25,
31).
The nature, history, and character of
allochthonous terranes along the Pacific
margin are best understood in the north-
ern cordillera of western North America
(1), particularly in southern Alaska and
British Columbia (Fig. 7). Among the
best known allochthonous terranes in
this region that may have been oceanic
plateaus are Wrangellia (37) and Cache
Creek (38).
Fig. 5. Tectonic elements
along the western South and
Central America consumption
zone (15): trench, active vol-
canoes, and seismicity. Num-
bers are depths in kilometers
of the seismic planes. Arrows
show direction of motion of
oceanic plates. Several aseis-
mic ridges are presently collid-
ing with the continents, caus-
ing volcanic and seismic gaps
on land.
liD
Wrangellia terrane. Wrangellia is
characterized by an enormous carapace
of Middle(?) to Upper Triassic subaerial
basalt, locally attaining a thickness of
6000 m, that overlies an upper Paleozoic
volcanic arc assemblage with associated
Permian and Triassic sedimentary rocks.
Over the Triassic basalt is a thick car-
bonate sequence of Late Triassic age,
which commences with inner platform
limestone and dolomite and ends in ba-
sinal pelagic carbonates, siliceous argil-
lite, and carbonaceous shale. Since con-
tinentally derived clastic material is
wholly lacking in this sequence, deposi-
tion in an oceanic setting seems man-
datory.
Two broad cycles of uplift and subsi-
dence are recorded in the upper Paleozo-
ic and lower Mesozoic stratigraphy of
Wrangellia. The first is represented by
10
f)
o
\600
10
100 90 80
shallow-water carbonate rocks with as-
sociated fossiliferous sandstone, shale,
and conglomerate of Permian age,
capped by a thin sequence of radiolarian
chert that ranges from Permian to Middle
Triassic in age. The subsidence may
have been caused by cooling of the un-
derlying upper Paleozoic volcanic arc.
Rapid uplift is recorded by the sudden
appearance above the cherty rocks of
Triassic amygdaloidal basalt (iocally pil-
lowed at the base). This basalt erupted
throughout Wrangellia with a total vol-
ume in the range of 100 to 200 km
3
, and
probably represents rifting related to the
commencement of northward movement
of the Wrangellian block from southern
paleolatitudes (32). A second broad epi-
sode of subsidence is recorded by the
thick Upper Triassic inner platform to
basinal deposits that overlie the basalt.
Table I. Paleomagnetic evidence for large-scale migration of allochthonous terranes now embedded in the Pacific margins.
Region
East Siberia
Northeast Siberia
Western Canada
and southern
Alaska
Japan
California
3 JULY 1981
Position and age
Sikhote-Alin. Since the Permian, 40 poleward motion relative to Siberian platform. Since the
Triassic, 20 poleward motion. Collision by Cretaceous. Moved 2000 kilometers in ~ 100
million years or at an average rate of 20 millimeters per year.
Kolyma block. Since the Permian, 20 poleward motion relative to Siberian platform. Since
the Triassic, 13 poleward motion. Collision with Siberian platform by Cretaceous.
Wrangellia terrane. Formed either at 18N or 18S of equator in late Triassic; the southern
latitude is more likely. Accreted by end of Cretaceous. Probably 6000 kilometers of north-
ward displacement in 130 million years for average rate of 46 millimeters per year.
Stikine terrane. Northward displacement (13) since late Jurassic.
Alaska Peninsula-Shumagin Islands. Northward movement ( ~ 50) since the Cretaceous or
~ 50 millimeters per year.
Inner belt of central Japan. In the Permian was situated near the paleoequator and was ac-
creted to the Asian mainland by the late Mesozoic.
Franciscan. Northward movement ( ~ 20) of seamounts relative to North America. Accreted
in Franciscan melange in late Cretaceous or early Cenozoic.
Refer-
ence
(75, 76)
(32, 76)
(77)
(78)
(79)
(80)
---------------------------------------------------------------
51
This subsidence appears to follow a post-
rifting cooling curve similar to that of
rifted continental margins (39).
Cache Creek terrane. The Cache
Creek terrane extends throughout much
of the central part of the Canadian cordil-
lera in a setting well inland from the
present continental margin (Fig. 7). The
presence within this terrane of non-
North American Permian fusulinids be-
longing to the Tethyan faunal province
led to the recognition of this terrane as
allochthonous (40).
Characteristic rocks of the Cache
Creek terrane (41) are mafic and ultra-
mafic rocks (ophiolites), chert, argillite,
pelite, volcanic sandstone and tuff, and
thick piles offossiliferous carbonate with
minor lenses of basic volcanics. The
assemblage of rocks in the Cache Creek
terrane may represent deposition in an
oceanic environment (38) in which local-
ly thick (2000 m) carbonate banks
formed plateau-like buildups that persist-
ed from early Carboniferous until Late
Permian time. Shallow-water fossils oc-
curring throughout these banks indicate
very slow progressive subsidence of the
basement, with final termination of car-
B
bonate deposition in Triassic time. Co-
eval deposition of deep-water rocks in
the Cache Creek terrane is demonstrated
by the presence of radiolarian cherts
ranging in age from Mississippian to
Triassic. Slide blocks of shallow-water
limestone occur locally in these deeper
water facies (41). Possible modern ana-
logs of the Cache Creek limestone banks
are large atolls or the Bahama Banks
(38).
Paleomagnetic data are not yet avail-
able to determine the paleolatitude of
formation of the Cache Creek terrane,
but paleobiogeographic analysis sup-
ports minimum movements of 30 north-
ward for the Tethyan fusulinid-bearing
limestones (42).
Continental Accretion
The role of continental collision in
orogenesis has long been recognized for
an area such as the Himalayas, where
two land masses are juxtaposed along a
major suture zone. The role of accretion-
ary tectonics in a mountain belt such as
the cordillera of western North America,
Fig. 6. Conceptual
model for the forma-
tion of the Aleutian
Ridge (25). Bowers
Ridge and U mnak
Plateau are thought to
have come from the
south with the Kula
plate and Shirshov
Ridge to have formed
in place. (A) Late
Mesozoic time and
(B) early Tertiary.
p ~ i ~ : i ~ ~ o n of plate
52
........... Active subductkln
zone
~ ~ Former subduction
zone
- Transform fault
which directly faces a vast open ocean,
has only recently been recognized (35,
43-46). Although the North American
cordillera and other mountain belts of
the Pacific rim are widely recognized as
products of accretion, little is under-
stood about the processes involved or
even the structures produced during
incorporation of allochthonous terranes
into the continental structure. Under-
standing this mechanism of continental
growth remains one of the fundamental
problems in geodynamics.
The gross structural relations in vari-
ous parts of the cordillera indicate that
thrust faulting played a dominant role in
the historical development of the en-
tire tectonic collage. A local structural
style consisting of large-scale imbricat-
ed thrust sheets is well documented in
southern Alaska (47-49), British Colum-
bia (50, 51), northwestern Washington
(52-54), the Klamath Mountains of
northern California (55-58), central Cali-
fornia (59--61), and southern California
(62). The amount of local differential
movement along some of the anastomos-
ing thrust faults must certainly exceed
several hundred kilometers, and move-
ments taken up within the entire accre-
tionary belt may well exceed 10,000 km.
Most of the accreted material consists of
blocks of thickened crust including arcs,
seamounts, oceanic crust overlain by
thick accumulations of sediments, pla-
teaus, and continental fragments. Ocean-
ic crust with thin sedimentary cover has
mostly disappeared. Thus, subduction
and underthrusting play key roles in this
accretionary process, but how and
where the thin thrust sheets are peeled
offfrom their lower crustal substrata and
are emplaced at supracrustal levels is not
apparent. Many nappe-like bodies have
been thrust onto the continental margin
or onto previously accreted terranes, but
there is no evidence of concomitant arc
and subduction activity. This makes it
exceedingly difficult to apply simple
plate tectonic models to any specific
locale as causes and effects within the
entire system cannot yet be related.
Subduction Versus Collision Orogeny
For several decades, two types of
mechanisms, collision and subduction,
have been suggested for orogenesis. Col-
lision is typified in the Alpine and Hima-
laya mountain chains. Subduction is typ-
ified by many of the circum-Pacific
mountain chains, traditionally those in
Alaska, western North America, east
Siberia, and particularly the Andes in
South America.
SCIENCE, VOL. 213
It now appears to us that Andean
orogeny, in the sense of subduction of
normal oceanic crust beneath a conti-
nent, has not been the underlying tecton-
ic process at the northern Pacific rim.
Wherever enough structural, stratigraph-
ic, and paleomagnetic data have become
available, it appears that allochthonous
terranes are commonly present and that
orogeny is intimately linked with the
incorporation of these terranes. Al-
though the occasional arc-continent col-
lision orogeny has been recognized (63),
we suggest that a very large fraction of
all orogenic episodes are the result of
collision. Aside from the Andean chain
and perhaps the Sunda arc, almost all
orogenies, or at least their deformation
phases, are associated with such colli-
sions. Little or no orogenic deformation
occurs where only pure subduction of
simple oceanic crust has taken place.
It is possible that allochthonous ter-
ranes actually played a role in the Andes
comparable to their role in other parts of
the Pacific rim, but the geological data
for western South America are insuffi-
cient to determine whether the required
allochthonous terranes are present. Nev-
ertheless, evidence is accumulating that
the orogenic history of the Andes is not
as simple as that expected from simple
subduction.
Several features stand out in particular
(64): (i) The Andes are made up of sever-
al tectonically and stratigraphically dis-
tinct geological assemblages, possibly al-
lochthonous terranes, which have been
welded together over a wide range of
geological time (65). (ii) Many Paleozoic
and early Mesozoic structures run
obliquely to the overall north-south
structural trend of the Andes, including
regions with deformation that penetrated
into continental basement rocks of late
Paleozoic age. (iii) Deep crustal fractures
provide sharp boundaries between sec-
tions of the Andes. Some of these sec-
tions differ from one another in their
geological history and rock types. In the
northern Andes, these sections are char-
acterized by rocks with oceanic affini-
ties, whereas from Peru south, rocks
have mostly continental affinities. (iv)
Prominent and extensive continental
basement rocks are eXposed along the
western coast from Tierra del Fuego to
Peru, with ages ranging from 1.8 billion
to 300 million years. These basement
rocks have been greatly deformed in
Paleozoic and Precambrian times, but
only mildly since. (v) Many investigators
suggest that continental sources to the
west of the Andes fed voluminous late
Paleozoic and early Mesozoic conglom-
erates and sandstones now found in the
3 JULY 1981
Andean chain (66, 67). Arc terranes in-
corporated from the west have also been
invoked by geophysicists (68) to explain
the presence of old continentai basement
off the Peru coast.
We believe that these general observa-
tions, while lacking in detail, open the
possibility that allochthonous terranes
have played a major role in the Andean
orogenic belt. It is possible that the con-
cept of Andean-type orogeny (orogeny
produced by subduction of oceanic crust
beneath a continent) is invalid. In other
words, it may well be that only one type
of process is responsible for orogenic
deformation, namely, collision. To test
this hypothesis, key areas in the Andes
must be studied to determine whether
major allochthonous terranes are embed-
ded in the Andean belt.
Conclusion
The role of allochthonous terranes in
continental accretion and mountain
building is becoming apparent. Some of
these terranes were probably oceanic
plateaus at one time during their past.
Many of the hundred or so plateaus in
today's oceans are fated to be incorpo-
rated at active continental margins, as
many must have been in the past. The
immediate effects of the accretion of
plateaus include the control of volcanic
activity and deep seismicity, trapping of
oceanic crust, and shifting of subduction
zones.
More lasting effects of the accretion of
plateaus are the growth of continental
crust and deformation in orogenic belts.
We suggest that all orogenic belts, even
Principal terranes
Alaska
NS North Slope
Kv Kagvlk
En Endicott
A Auby
Sp Seaward Peninsula
I I"noko
NF Nixon Fork
PM Plngston and McKinley
YT Yukon-Tanana
CI Chulitna
P Peninsula r
W Wrangellia
Cg Chugach lind Prince William
TA Tracy Arm
T Taku
Ax Alexander
G Goodnews
Canada
Ch Cache Creek
St Stlklne
BA Bridge Alver
E Eastern assemblages
Washington. Oregon. and California
Ca Northern Cascades
SJ San Juan
o Olympic
S Siletzla
BL Blue Mountains
Trp Western Triassic and Paleozoic
of Klamath Mountains
KL Klamath Mountains
Fh Foothills belt
F Franciscan and Great Valley
C Calaveras
SI Northern Sierra
SG San Gabriel
Mo Mohave
Sa Sallnla
Or Orocopia
Nevada
S Sonomla
AM Roberts Mountains
GL Golconda
Mexico
B Baja
V Vizcaino
Fig. 7. Map showing the distribution of principal tectonostratigraphic terranes in North America
(1). The extent of the craton is shown by pattern. The barbed line marks the eastern limit of
cordilleran Mesozoic-Cenozoic deformation. Possible examples of oceanic plateaus are shown
by horizontal lines.
53
those classified as the subduction type,
may in fact be the result of collisions-
collisions not with major continents but
with oceanic plateaus, whose origins in-
clude extinct arcs, submerged continen-
tal fragments, clusters of seamounts, and
hot spot traces.
The link between allochthonous ter-
ranes on land and migrating plateaus in
the oceans provides a new way to relate
land geology to the marine-derived con-
cept of plate tectonics. Instead of envi-
sioning vast oceans in the past underlain
by simple ocean floor, we must think in
terms of a more complex oceanic geolo-
gy in which many plateaus with different
origins were embedded in ancient ocean-
ic plates, just as they are today.
References and Notes
I. P. J. Coney, D. L. Jones, J. W. H. Monger,
Nature (London), 288, 329 (1980).
2. D. E. Karig, Geol. Soc. Am. Bull. 83, 1057
(1972).
3. K. O. Emery and B. J. Skinner, Mar. Min. 1,
(1977).
4. E. L. Winterer, Geophys. Monogr. Am.
Geophys. Union 19, 269 (1976).
5. H. W. Menard, Marine Geology of the Pacific
(McGrawHill, New York, 1964).
6. A. K. Cooper and Z. BenAvraham, Eos 60, 950
(1979).
7. L. W. Kroenke, "Geology of the Ontong Java
Plaieau," Hawaii Univ. Inst. Geophys. Rep.
72:5 (1972).
8. K. O. Emery, personal communication.
9. R. Houtz, personal communication.
10. R. Hey, Geol. Soc. Am. Bull. 88, 1404 (1977).
II. J. E. Andrews et al., Init. Rep. Deep Sea Drill.
Pro}. 30 (1975).
12. S. R. Hammond et al., ibid., p. 415.
13. P. R. Vogt, A. Lowrie, D. R. Bracey, R. N.
Hey, Geol. Soc. Am. Spec. Pap. 1972 (1976).
14. E. Bonatti, C. G. A. Harrison, D. E. Fisher, J.
Hannorez, G. Schilling, J. J. Stipp, M. Zentilli,
1. Geophys. Res. 82,2457 (1977).
15. A. Nur and Z. Ben-Avraham, Geol. Soc. Am.
Mem., in press.
16. M. Barazangi and B. Isacks, Geology 4, 696
(1976).
17. ____ , Geophys. 1. R. Astron. Soc. 57, 537
(1979).
18. B. L. Isacks and M. Barazangi, in Island Arcs,
Deep Sea Trenches and Back Arc Basins, M.
Talwani and W. C. Pitman, Eds. (American
Geophysical Union, Washington, D.C., 1977),
pp.99-114.
54
19. H. K. Eisler and H. Kanamori, Eos 60, 878
(1979).
20. J. F. Dewey, Sci. Am. 226, 56 (May 1972).
21. A. Hasegawa and I. S. Sacks, Eos 60, 876
(1979).
22. 1. Kelleher and W. McCann, in Island Arcs,
Deep Sea Trenches and Back Arc Basins, M.
Talwani and W. C. Pitman, Eds. (American
Geophysical Union, Washington, D.C., 1977),
pp. 115-122.
23. H. Kanamori, 1. Geophys. Res. 82, 2981 (1977).
24. W. R. Dickinson, 1. Phys. Earth 26, SI (1978).
25. Z. BenAvraham, A. K. Cooper, D. W. Scholl,
Geol. Soc. Am. Abstr. Programs 21,96 (1980).
26. A. K. Cooper, M. S. Marlow, D. W. Scholl, 1.
Geophys. Res. 81, 1916 (1976).
27. G. B. Dalrymple, D. A. Clague, M. A. Lan-
phere, Earth Planet. Sci. Lett. 37, 107 (1977).
28. W. J. Ludwig et al., 1. Geophys. Res. 76,6350
(1971).
29. A. K. Cooper, D. W. Scholl, T. L. Vallier, E.
W. Scott, U.S. Geol. Surv. Open-File Rep. 80-
246 (1980).
30. S. Uyeda and A. Miyashiro, Geol. Soc. Am.
Bull. 85, 1159 (1974).
31. T. Matsuda, 1. Phys. Earth 26, S409 (1978).
32. J. W. Hillhouse, Can. 1. Earth Sci. 14, 2578
(1977).
33. A. Cox, Geol. Assoc. Can. Spec. Pap. 20, 305
(1980).
34. R. C. Speed, 1. Geol. 87, 279 (1979).
35. J. W. H. Monger and R. A. Price, Can. 1. Earth
Sci. 16, 770 (1979).
36. A. Nur and Z. Ben-Avraham, 1. Phys. Earth.
26, S21 (1978).
37. D. L. Jones, N. J. Silberling, J. W. Hillhouse,
Can. 1. Earth Sci. 14, 2565 (1977).
38. J. W. H. Monger, ibid., p. 1832.
39. N. H. Sleep, Geophys. 1. R. Astron. Soc. 24,
325 (1971).
40. J. W. H. Monger and C. A. Ross, Can. 1. Earth
Sci. 8, 259 (1971).
41. J. W. H. Monger, Geol. Surv. Can. Pap. 7447
(1975).
42. T. E. Yancey, in Historical Biography, Plate
Tectonics, and Its Changing Environment, J.
Gray and H. J. Boucot, Eds. (Oregon State
Univ. Press, Corvallis, 1979), p. 239.
43. J. Dercourt, Can. 1. Earth Sci. 9, 709 (1972).
44. G. A. Davis, J. W. H. Monger, B. C. Burchfiel,
in Mesozoic Paleogeography of the Western
United States, D. G. Howell and K. A. McDou-
gall, Eds. (Society of Economic Paleontologists
and Mineralogists, Tulsa, Okla., 1978), p. I.
45. W. Hamilton, in ibid., p. 33.
46. D. L. Jones and N. J. Silberling, U.S. Geol.
Surv. Open-File Rep. 79-1200 (1979).
47. B. Csejtey, Jr., U.S. Geol. Surv. Circ. 804-B
(1979), p. B90.
48. D. J. Jones, N. J. Silberling, B. Csejtey, Jr., W.
H. Nelson, C. D. Blome, Geol. Surv. Prof. Pap.
1I2I-A (1980).
49. H. C. Berg, D. L. Jones, P. J. Coney, U.S.
Geol. Surv. Open-File Rep. 78-/085 (1978).
50. D. J. Tempelman-Kluit, Geol. Surv. Can. Pap.
79-]4 (1979), p. I.
51. W. B. Travers and J. H. Ladd, Geol. Soc. Am.
Abstr. Programs 11, 529 (1979).
52. P. Misch, Can. Inst. Min. Metall. Spec. Vol. 8,
101 (1966).
53. J. T. Whetten et al., i:1 Mesozoic Paleogeogra-
phy of the Western United States, D. G. Howell
and K. A. McDougall, Eds. (Society of Eco-
nomic Paleontologists and Mineralogists, Tulsa,
Okla., 1978), p. 117.
54. J. T. Whetten et al., Geol. Soc. Am. Bull. 91,
359 (1980).
55. W. P. Irwin, Calif. Viv. Mines Geol. Bull. 190,
19 (1966).
56. ___ , U.S. Geol. Surv. Prof. Pap. 800-C
(1972), p. C103.
57. G. A. Davis, Geol. Soc. Am. Bull. 79, 911
(1968). ..
58. __ , ibid. 80, 1095 (1969).
59. E. H. Bailey, M. C. Blake, Jr., D. L. Jones,
U.S. Geol. Surv. Prof. Pap. 700-C (1970), p.
C70.
60. J. Suppe and K. A. Foland, in Mesozoic Paleo-
geography of the Western United States, D. G.
Howell and K. A. McDougall, Eds. (Society of
Economic Paleontologists and Mineralogists,
Tulsa, Okla., 1978), p. 431.
61. J. Suppe, Geol. Soc. Am. Bull. 90 (part I), 327
(1979).
62. G. Haxel and J. Dillon, in Mesozoic Paleogeog-
raphy of the Western United States, D. G.
Howell and K. A. McDougall, Eds. (Society of
Economic Paleontologists and Mineralogists,
Tulsa, Okla., 1978), p. 453. .
63. D. Roeder, Rev. Geophys. Space Phys. 17, 1098
(1979).
64. W. Zeil, The Andes, A Geological Review
(Borntraeger, Berlin, 1979).
65. A. Gansser, 1. Geol. Soc. London 129, 93
(1973).
66. H. Miller, Geotektonische Forsch. 36, I (1970).
67. P. Isaacson, Geol. Soc. Am. Bull. 86, 39 (1975).
68. D. E. James, ibid. 82, 3326 (1971).
69. A. S.Furumoto, W. A. Wiebenga, J. P. Webb,
G. H. Sutton, Tectonophysics 20, 153 (1973).
70. G. Palmason, VisindaJelag lsi. 40, I (1971) (en-
tire volume).
71. A. S. Laughton, D. H. Matthews, R. L. Fisher,
The Sea, A. E. Maxwell, Ed. (Wiley, New
York, 1970), vol. 4.
72. H. Narain, Tectonophysics 20, 249 (1973).
73. D. E. James, 1. Geophys. Res. 76,3246 (1971).
74. J. B. Minster and T. H. Jordan, ibid. 83, 5331
(1978).
75. M. W. McElhinny, Paleomagnetism and Plate
Tectonics (Cambridge Univ. Press, Cambridge,
1973).
76. M. W. McElhinny, The Western Pacific: Island
Arcs, Marginal Seas, Geochemistry, P. J. Cole-
man, Ed. (Univ. of Western Australia Press,
Nedlands, 1973), p. 407.
77. E. Irving, J. W. H. Monger, R. W. Yole, Geol.
Assoc. Can. Spec. Pap. 20 (1980), p. 441.
78. D. B. Stone and D. R. Packer, Geol. Soc. Am.
Bull. 90, 545 (1979).
79. I. Hattori and K. Hirooka, Tectonophysics 57,
211 (1979).
80. C. S. Gromme and H. J. Gluskoter,l. Geophys.
Res. 73, 74 (1965).
81. We thank S. McGeary and M. Martin for their
help in preparing the figures. Supported in part
by the National Science Foundation.
SCIENCE, VOL. 213